Asymptote
Banned Idiot
^ Very informative and interesting, but doesn't really answer my question.
I'll put it another way (not demanding you answer it of course, just disputing a point that asymptote made a few pages back).
Let's say the trailing edge of the J-20s canard were aligned with the main wing's trailing edge on the same side. Does that produce a massively different radar return than if the trailing edge were aligned diagonally?
^ Red line shows current "diagonal" alignment, blue lines show the a hypothetical modified canard with a trailing edge which aligns on the same side with the wing behind it rather than the wing opposite/behind it as is the J-20's current "diagonal" case.
It probably does. The Korean stealth project "KFX" is design just like above - they probably think its better to align it on the same side than diagonally.
Korean KFX
Remember, there is the main aircraft body in middle so aligning diagonally seems pointless if the other side get shielded by the main airframe body. So its better and more logical to just align on the same side, when the radar wave hits from one direction (let's say 45 degree off the aircraft) all the edges from one side will be aligned.